Last post Sep 15, 2011 11:40 PM by Decker Dong - MSFT
Sep 14, 2011 12:17 PM|dotnetdevdallas|LINK
Iam new to ssis stuff. the thing is i developed a windows service which runs every 1 hr and write some formatted data to a text file from a text file.
now the architecture team is more concerned about memory issues if we deploy windows service and they want it to run as an ssis package.
so can somebody shed some light on how we can build a package which takes dta from a text file(source ) and generate a new text file with formatted data.
any kind of help wud be appeciated to kick startof....
Sep 14, 2011 12:57 PM|jpazgier|LINK
Are you sure you would like to use SSIS package? Maybe if it is SQL 2005 or higher
you can run CLR stored procedure to do the work?
Be aware that you can also load text file into SQL server directly using Bulk
insert and save it back with simple CLR save to file procedure.
If you have SQL 2008 you can also use power shell step in SQL agent and do all
work using it.
SSIS package is a very good and powerful tool but I believe it is for users who
started they programming work or if you really would like to use tools
available in it and will be hard or time consuming to do processing different
In your case if service is already written in .Net language (c#, VB) you can put your code
directly into Script step in the SSIS package.
Another solution is to convert your service process to standalone batch application and run it in DOS
batch step in SQl agent job
Sep 14, 2011 02:21 PM|dotnetdevdallas|LINK
Hi Thanks for the reply!!
I already have the code written in csharp...so i wud like to opt for ssis package...as u said..are there any particular examples on this how to start the etl process in this case...
Sep 14, 2011 02:45 PM|konanki|LINK
using windows service
Sep 15, 2011 11:40 PM|decker dong - msft|LINK
can somebody shed some light on how we can build a package which takes dta from a text file(source ) and generate a new text file with formatted data.
In fact I think you can try the following two detailled samples: